3D character
Keen Software House Prague
themes
science, artificial intelligence, space exploration, biodiversity
reading time
7 minutes
A small, inconspicuous looking thing pinned to a yellowed page. I am looking at some kind of wasp that was collected in the 1960s by two entomologists named M. and J. Wasbauer.
Little pieces of paper alongside the specimen contain scribbles in black ink. Roman numerals refer to the number of the month it was collected in. Apparently, this is standard practice among insect scientists.
I carefully type in when and where in California the bug was collected. In a box for additional information, I mention the trap used to catch it. Wondering what the scientists had looked like and what kind of landscape they had found themselves in, my mind wanders a little. Outside my window, Brussels looks lacklustre under the grey February sky. But I can smell the grass that has dried to a crisp somewhere in the USA. Scents of the faithful day that the animal on my screen lost its life to science.
The dead bug photos on my computer were taken for the California Terrestrial Arthropod Database Project, CalBug for short. The project aspires to digitise California’s eight major terrestrial arthropod collections for science. Data that, according to their website, can be used “to study biogeographic patterns, spread of invasive species, and responses to land use, climate, and other environmental changes.”
People transcribe label data for CalBug through citizen science platform Zooniverse. Volunteers can engage in different scientific disciplines, from astrophysics to art history, climate science, and indeed, entomology. You don’t need a PhD to participate. Some free time and an internet connection do the trick.
WWW
Zooniverse has proven particularly useful for processing big batches of data. The CalBug “Notes From Nature” project aims to digitise 74,396 subjects. A feat that would take a trained individual 78 days of straight work and no sleep. Or the equivalent of 233 eight-hour working days if this person would never get distracted and kept a steady pace throughout. In reality, it would take much longer. My most productive session reaped 39 completed specimens in about one and a half hours—about a third slower than my estimated speed of a trained pro.
Digitising 74,396 specimens would be a near impossible—and extremely expensive—task for one person or small team. So, what if we could get the work done, but for free? While engaging non-scientists in important research for fun. Sounds good?
So, what if we could get the work done, but for free?
Who knows, maybe we’ll find something cool. Like a previously unknown astronomical object discovered by a Dutch schoolteacher in the Netherlands. Hanny van Arkel was volunteering for Galaxy Zoo, Zooniverse’s predecessor, when she spotted a strange-looking blob while sorting through images of astronomical objects in 2007.
Scientists were already training computers to recognise astronomical objects, but it remained essential to check using human eyes and brains. To train an AI to do the work, you need to feed it as many human-labelled examples as possible. But what if something has never been seen before? An AI trained to distinguish cats and dogs would be clueless when presented with a cactus. Let alone if an AI is asked to interpret an object that humans didn’t know existed.
Hanny was curious about this weird shape in space, so she flagged it. That’s all it took for her name to go down in history as the discoverer of a new astronomical phenomenon, Wikipedia page and all.
One Redditor tells the story of when they identified a supernova and their Zooniverse handle was mentioned in a research paper. Afterwards, they even went on TV to chat about it to British physicist Brian Cox. Another Redditor recalled how they participated in a NASA project identifying objects on the surface of Mars to train the Martian rovers to drive themselves.
WWW
Scientists estimate that they have described 10% of the biodiversity on Earth. A 2012 study found that the average “new thing to science” has a shelf life of 21 years between “being discovered” and being described. In other words: countless new species already grace the shelves of natural history museums, but we just haven’t gotten around to them. Taxonomists—experts in identifying a certain group of organisms—are a rare breed. While the importance of biodiversity becomes increasingly clear in the face of biodiversity decline and climate change, we are struggling to figure out what we need to protect.
We are struggling to figure out what we need to protect
So, I click the save button and move on to the next wasp. Equally nondescript and not very waspy, if I’m honest. Turns out that most wasps don’t really look like wasps at all. That is, like the infamous black-and-yellow creatures who love to visit your summer picnic. Beetles were long thought to be the most numerous order in the animal kingdom, but—unfortunately for the beetles—this is not the case. Scientists now believe that Hymenoptera (bees, wasps, and ants) come in first, mostly due to wasp diversity.
I come across another specimen collected by someone called Wasbauer, the name I ran into earlier. It was collected in 2006, almost half a century after the other specimen I labelled. Was this the same person, or a descendant, perhaps? And what about the Wasbauers who collected the earlier specimen in the 60s. Were they siblings, relatives, or a couple?
Spurred on by curiosity, I decide to do some online sleuthing. There can only be so many Wasbauers with a special affinity for wasps, I think to myself. The first hit looks promising: “Marius Wasbauer's Massive Wasp Specimen Collection Donated to Bohart Museum”. I click the link.
“The UC Davis Bohart Museum of Entomology recently received a massive donation of more than 50,000 aculeate wasp specimens, primarily spider wasps, from the estate of renowned wasp expert Marius Wasbauer,” the article reads. Marius Wasbauer died a few months earlier, in the spring of 2021. His legacy consists of a mountain of research and 180 wooden drawers of specimens collected from the 60s well into the 2010s. The reason I am labelling all these wasps is that Marius Wasbauer gifted his collection to a museum. And this collection is now being digitised by Zooniverse volunteers for CalBug.
A legacy consisting of a mountain of research and 180 wooden drawers of specimens
I’m starting to connect the dots. It was the same person who collected the 60s and the 2006 specimens I labelled. A photo of Marius Wasbauer in 2017, butterfly net in hand and a camera around his neck, appears alongside the article. The woman next to him is looking at a bug in a red-capped vial. “Marius Wasbauer and his wife Joanne, at the Bohart Museum Bioblitz to Belize in 2017,” the caption says.
My eyes widen. Joanne is the second Wasbauer mentioned on the 1960s label. Joanne and Marius caught that little beast in their thirties, during the days of John F. Kennedy, Jimi Hendrix and the Apollo 11 mission to the Moon. They went into the field together then, and still did in 2017, when their photo was taken.
WWW
Although volunteers can interact with each other and with scientists through a message board called “Zooniverse Talk”, community building on the platform itself seems to be lagging behind. People discuss the project more widely on social media and online citizen-science related communities.
But maybe an active message board is not a requirement for a thriving online community. I have told people that I am a “citizen scientist” (with air quotes) at parties before. Stumbling on the Wasbauers made me feel connected to a young couple of entomologists on their quest for Californian wasps back in the 60s. And hearing about people’s discoveries online inspires me to keep participating. Making a contribution to science from my laptop makes me part of network of thousands who are doing the same. If science is a group effort, doesn’t that make those who engage a part of the same community?
Thousands of eyes and ears give scientists millions of data points to play with
Other citizen science platforms, like iNaturalist and Observations.org, even manage to get people outside. Snapping pictures of whatever wild flora and fauna you encounter, AI identifies what species you are looking at. Simultaneously, the photo is logged in a database, automatically attaching the time and location. The interfaces of their apps look enticing, while challenges and fancy graphs invite people to go out and explore more. Enthusiasts meet each other online and in the field to go on expeditions and learn from each other.
Although observations made for iNaturalist of Observations.org are not directly tied to a research question or team like in the Zooniverse, thousands of eyes and ears give scientists millions of data points to play with. In an era of biodiversity decline, researchers need this information more than ever to establish baselines and tease out trends.
The time we spend online can fly by, especially when binge- or doomscrolling a tireless feed. Fortunately, citizen science platforms are equally adept at making time disappear. Only the hangover feels so much better. Citizen science taps into an endeavour that has given people a sense of wonder and purpose since the beginning of time—to get to know more about the world and the universe.